This invention relates generally to sensing systems utilizing cryogenic cooling and, more particularly, to a device including integrated optical waveguides for providing switching and modulation of optical power.
Optical detecting systems for sensing infrared radiation employ sensors which operate at optimal performance levels at cryogenic temperatures. As examples, mercury cadmium telluride photodiodes are usually maintained at 77.degree. K., while copper-doped germanium devices are normally operated at 4.degree. K. The sensing devices and their support electronics are typically housed in a cryostatically pumped Dewar in order to maintain these low temperatures.
Current sensing systems, as may be employed in a satellite system which scans portions of the heavens for infrared radiation, may include arrays having literally millions of photodiode sensors, thereby introducing a significant problem of coupling information between the sensors and their immediate support electronics, operating in a cryogenic region, and the data processing electronics, operating at a much higher temperature. The temperature gradient creates thermal and mechanical stresses upon any element physically coupling the two regions.
Cryogenic signal handling techniques in current use often compromise the sensor's front-end in terms of electrical performance, cost and physical space consumed. The primary functions involved in any infrared staring or scanning sensor are signal acquisition, transduction, routing and processing. Each piece of raw data, whether discrete analog pixels or continuous, must be carefully handled and communicated to other points.
The degree by which a sensor system performs its primary functions can be measured in terms of noise generation, channel-to-channel crosstalk, power dissipation, physical size and weight, and environmental robustness. Cryogenically cooled sensors have the additional problem that they must be operated at very low temperatures while retaining an optimally balanced performance as determined by these operational standards.
When employing conventional electronic devices, typically CMOS, to process the sensor information, channel-to-channel crosstalk in the electronics partially negates the noise reduction benefits afforded by cooling the sensor's front-end. Heat generation from the CMOS further aggravates the sensor's noise generation at the front-end. This puts an added load on the sensor's cryostat by increasing its size, weight and operating power requirements. Additionally, conventional electronic devices are known to have gain and DC operating point instabilities at temperatures in the 77.degree. K. region. Finally, real estate is generally at a premium on the cryogenic front-end and any reduction in componentry reaps an operational advantage.
The primary problems associated with conventional CMOS signal handling in the cryogenic region include electronically generated heat which raises the sensor's noise floor, compromised performance of CMOS circuitry due to cryogenically induced gain and DC instabilities, electromagnetic interference between the low level photodiode output signals and supporting electronic control signals, and channel-to-channel crosstalk due to individual channel leakage and multiplexing control signals. These final two problems contribute significantly to a lowering of the effective signal-to-noise ratio for any given sensor channel beyond the infrared transduction noise.
Another common problem with conventional multiplexing of mercury cadmium telluride diode signals is the increased switching speeds required with a large number of channels. These increased switching speeds increase crosstalk among channels and require increasingly smaller sampling times to allow for transient settling at the end of the cascaded multiplexer chain. This puts design pressure on the A/D conversion which follows the sensor stage and on the and signal processing requirements for real-time systems.
Since the early 1980's, the industry has studied CMOS technology operating at cryogenic temperatures to process focal plane array data before the signal-to-noise ratio could be corrupted. It has been found that the operation of CMOS devices at liquid nitrogen temperatures produces "freezout" effects in depletion and enhancement-mode CMOS devices. One such "freezout" effect is a cryogenically stressed device degradation referred to as hot-electron device degradation, which creates transistor threshold changes and greatly diminished transconductance. Together, these degradations create instabilities in gain and DC levels.
Another problem associated with cryogenic operation is increased interconnection resistance among devices that are fabricated at submicron scale. This phenomenon creates signal bandwidth limitations.
Finally, a physical interconnection between the cryogenic and noncryogenic regions introduces a thermal leakage path when electrical connections are anchored to a backplane. This supplies additional heat to the sensor and necessitates greater cryostat pumping capability.